Active power distribution network optimal scheduling method and system considering multi-agent cooperative game

Abstract

The invention discloses an active power distribution network optimal scheduling method and system considering a multi-agent cooperative game, and relates to the technical field of power distribution network optimal scheduling, and the method comprises the steps: considering the interaction among multiple market agents in an active power distribution network, and constructing a multi-agent cooperative transaction architecture; based on a multi-subject collaborative transaction architecture, taking cost minimization as a target, and combining constraint conditions to build an operation and income model of each subject; wherein a distributed energy storage operator is used as one of main bodies, and the operation of distributed energy storage is restrained by using a distributed energy storage distribution strategy based on distributed energy storage SOC balance; according to the Nash bargaining theory, a multi-subject operation and income model is integrated, and a multi-subject cooperative game model is constructed; and solving the multi-agent cooperative game model based on the actual energy demand quantity of the user side to obtain an optimal scheduling strategy of the energy in the active power distribution network, and performing energy scheduling according to the scheduling strategy. The charging and discharging power balance between the DES units can be realized, and the economic benefit is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network optimal scheduling, and particularly to an active distribution network optimal scheduling method and system considering multi-agent cooperative game. Background Art

[0002] The statements in this part only provide background technical information related to the present invention and do not necessarily constitute prior art.

[0003] Under the background of the current increasingly severe energy crisis and environmental pollution, the rapid development of renewable energy generation technology has enabled it to be rapidly popularized globally and widely connected to the active distribution network (ADN) in the form of distributed generation (DG). Under the background of the electricity market reform, power users are encouraged to install distributed generation on the user side, which means that traditional power users can not only consume electric energy but also produce electric energy to participate in the electricity market transaction. Therefore, the power supply mode in which the ownership of distributed resources belongs to the distribution network has been changed. Compared with the traditional scheduling method, the coordinated optimal scheduling of the ADN containing multiple distributed resources can effectively integrate renewable energy (RES) while ensuring the safety, stability, and economic operation of the system. For example, existing research has proposed a multi-stage stochastic programming model for the day-ahead optimal scheduling of the active distribution network, enabling the scheduling decision to more accurately cope with the uncertainty of renewable energy; proposed applying the second-order cone relaxation technology to the day-ahead multi-period optimal model of the active distribution network to achieve the effective solution of the model; established a dynamic optimal power flow model on multiple time scales with the goal of maximizing the energy output of renewable energy generation, and reduced the curtailment of wind and light by coordinating the control of energy storage and flexible loads. The above research regards distributed resources as schedulable units, but does not consider that in the market-oriented situation, distributed resources are constructed and managed by operators, forming independent entities such as distributed energy storage operators (DESO), load aggregators (LA), etc., and each entity can participate in the optimal scheduling according to its own interest demands. Therefore, the problem of multi-agent collaboration still needs to be solved at present.

[0004] In addition, research on the participation of distributed energy storage (DES) in the optimal scheduling of ADN has been widely carried out. With a large amount of RES connected to the ADN through DG, the strong randomness and intermittency of its output result in large fluctuations in grid-connected power, leading to problems such as increased voltage fluctuations and degraded power quality in the system. Distributed energy storage has fast charge and discharge characteristics and can achieve energy transfer in time and space, which is an effective means to solve the power fluctuations in distribution network systems with distributed renewable energy. For example, existing research has proposed a novel battery charge and discharge control strategy for distribution systems to optimize energy trading costs; to alleviate imbalance problems, the power flexibility of DES is used to provide support to the upper-level grid at the point of common coupling, etc. Although existing research has effectively improved the operational stability of the system through DES scheduling, considering the inconsistent charge and discharge levels among distributed energy storages, some DESs will be overcharged and over-discharged, thus affecting the lifespan of the energy storage. Existing research has not taken into account the balance of SOC (State of Charge) among distributed energy storages, which means that in the actual operation process, there may be a serious inconsistency in the state of charge among energy storages, resulting in overcharging and over-discharging of some energy storages, and further causing problems such as low utilization rate of battery capacity and shortened service life. Summary of the Invention

[0005] To solve the deficiencies of the above-mentioned prior art, the present invention provides an active distribution network optimal scheduling method and system considering multi-agent cooperative game, which performs collaborative optimal scheduling by simultaneously considering the SOC balance of DES and the interests of multiple agents. That is, on the one hand, a DES power distribution strategy based on SOC balance is designed, dividing DES into a priority charging group and a priority discharging group, and through double-layer allocation between and within groups, the balance of charge and discharge power among DES units is achieved, reducing lifespan loss; on the other hand, a trading framework describing the economic interaction among the distribution system operator (DSO), distributed energy storage operator (DESO), and load aggregator (LA) during the energy trading process in the ADN is proposed, and then the Nash bargaining theory is introduced to establish a multi-agent cooperative game model to effectively improve the economic benefits of all participating entities.

[0006] In the first aspect, the present invention provides an active distribution network optimal scheduling method considering multi-agent cooperative game.

[0007] An active distribution network optimal scheduling method considering multi-agent cooperative game includes: Considering the interaction among multiple market players in the active distribution network, constructing a multi-agent collaborative trading architecture; Based on the multi-agent collaborative trading architecture, with the goal of minimizing cost and combining constraint conditions, building the operation and revenue model of each agent; among them, as one of the agents, the distributed energy storage operator uses a distributed energy storage allocation strategy based on the SOC balance of distributed energy storage to constrain the operation of distributed energy storage. According to the Nash bargaining theory, a multi-agent cooperation game model is constructed by integrating the operation and revenue models of multiple agents; Based on the actual energy demand on the user side, the multi-agent cooperation game model is solved to obtain an optimal energy scheduling strategy in the active distribution network, and energy scheduling is carried out according to the scheduling strategy.

[0008] Secondly, the present invention provides an active distribution network optimal scheduling system considering multi-agent cooperation game.

[0009] An active distribution network optimal scheduling system considering multi-agent cooperation game includes: An optimal scheduling model construction module, which is used to consider the interaction between multiple market players in the active distribution network, construct a multi-agent collaborative trading architecture; based on the multi-agent collaborative trading architecture, with the goal of minimizing cost and combining constraint conditions, build the operation and revenue model of each agent; among them, as one of the agents, the distributed energy storage operator uses a distributed energy storage allocation strategy based on the SOC balance of distributed energy storage to constrain the operation of distributed energy storage; according to the Nash bargaining theory, integrate the operation and revenue models of multiple agents to construct a multi-agent cooperation game model; A model solving and scheduling module, which is used to solve the multi-agent cooperation game model based on the actual energy demand on the user side, obtain an optimal energy scheduling strategy in the active distribution network, and carry out energy scheduling according to the scheduling strategy.

[0010] Thirdly, the present invention also provides an electronic device, including: a memory for storing executable instructions; a processor, when executing the executable instructions stored in the memory, implements the above-mentioned active distribution network optimal scheduling method considering multi-agent cooperation game.

[0011] Fourthly, the present invention also provides a computer-readable storage medium storing executable instructions, which are used to cause the processor to implement the above-mentioned active distribution network optimal scheduling method when executing the executable instructions.

[0012] Fifthly, the present invention also provides a computer program product, which includes executable instructions stored in a computer-readable storage medium; wherein, when the processor of the electronic device reads and executes the executable instructions from the computer-readable storage medium, it implements the above-mentioned active distribution network optimal scheduling method considering multi-agent cooperation game.

[0013] The above one or more technical solutions have the following beneficial effects: 1. The present invention provides an active distribution network optimal scheduling method and system considering multi-agent cooperative game. Through simultaneous consideration of the SOC balance of DES and the interests of multiple agents for collaborative optimal scheduling, on the one hand, a DES power distribution strategy based on SOC balance is designed. The DES is divided into a priority charging group and a priority discharging group. Through double-layer allocation between groups and within groups, the balance of charge and discharge power among DES units is achieved, reducing life loss. On the other hand, a trading framework describing the economic interaction among the distribution system operator (DSO), distributed energy storage operator (DESO), and load aggregator (LA) during the energy trading process in the ADN is proposed. Then, the Nash bargaining theory is introduced to establish a multi-agent cooperative game model to effectively improve the economic benefits of all participating entities.

[0014] 2. The present invention proposes a DES power distribution strategy based on SOC balance. The DES is divided into a priority charging group and a priority discharging group. The optimal allocation of the total power command between groups is determined through upper-layer allocation, and the relative balance of the SOC of distributed energy storage within the group is achieved through lower-layer allocation. Compared with traditional energy storage scheduling strategies, this strategy can ensure the dynamic relative balance of each energy storage SOC during scheduling, effectively improve the utilization rate of DES, and at the same time reduce the charge and discharge switching times of the energy storage, extending the service life to a certain extent and reducing the life loss of DES.

[0015] 3. The present invention constructs a multi-agent collaborative trading framework for distribution companies, load aggregators, and energy storage operators. This framework considers the interaction among DSO, DESO, and LA, defines trading mechanisms and flexible pricing strategies in different markets to promote the coordinated pricing of electricity transactions, and collaboratively optimizes electricity transactions to achieve win-win cooperation among multiple interest entities. On this basis, a multi-agent collaborative optimal scheduling model considering the SOC balance of distributed energy storage is constructed. This model takes maximizing social benefits as the optimization goal and realizes the fair distribution of interests based on the Nash bargaining theory. Compared with traditional scheduling methods, the present invention can make full use of the resources of each entity, greatly reduce the electricity purchase volume of the system from the upper-layer power grid, and achieve the maximization of social benefits.

[0016] Advantages of additional aspects of the present invention will be partly given in the following description, partly will become obvious from the following description, or will be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The specification drawings forming a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention.

[0018] Figure 1 It is the overall flowchart of the active distribution network optimal scheduling method considering multi-agent cooperative game according to the embodiment of the present invention; Figure 2 Schematic diagram of the multi - agent transaction framework in the embodiment of the present invention; Figure 3 Schematic diagram of 10 typical scenarios output by the fan in the embodiment of the present invention; Figure 4 Schematic diagram of the SOC change curve under Scheme 1 in the embodiment of the present invention; Figure 5 Schematic diagram of the SOC change curve under Scheme 2 in the embodiment of the present invention; Figure 6 Schematic diagram of the analysis result of the peak - shaving and valley - filling effect in the embodiment of the present invention. Detailed implementation manners

[0019] It should be noted that the following detailed description is exemplary only for describing the specific implementation manners, aiming to provide further explanation for the present invention, and is not intended to limit the exemplary embodiments of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. In addition, it should also be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0020] Embodiment 1 Considering that in the active distribution network (ADN), the state of charge (SOC) of distributed energy storage (DES) is unbalanced, and coupled with the intertwined interests of multiple stakeholders, it significantly increases the complexity of system coordination and scheduling. Therefore, this embodiment provides an optimal scheduling method for the active distribution network considering multi - agent cooperative game, which can simultaneously consider the SOC balance of DES and multi - agent interests for collaborative optimal scheduling. As Figure 1 shown, it specifically includes the following steps: Step S1: Considering the interaction between multiple market entities in the active distribution network, construct a multi - agent collaborative transaction architecture.

[0021] For the multi - agent collaborative transaction architecture studied in this embodiment, its market entities include the distribution system operator DSO, the load aggregator LA distributed in the distribution network, and the distributed energy storage operator DESO. The interconnection structure and interaction mechanism among multiple entities in the system are as Figure 2 shown. It must be emphasized that all distributed energy storage connected to the active distribution network is built by the energy storage operator, and it is allowed to serve both the distribution network and the load aggregator at the same time; in addition, since there is no direct line connection between multiple load aggregators and between the load aggregator and the distributed energy storage, power interaction needs to be carried out through the distribution network line. To ensure its own interests, the distribution company charges a transmission fee for these two trading forms.

[0022] Among them, the responsibilities of entities (i.e., each market entity) in the multi-agent trading structure are as follows: (1) In the active distribution network, controllable distributed power sources mainly based on gas turbines, conventional loads, and some reactive power compensation devices are equipped, mainly including static var compensators (SVCs), discrete capacitors (CBs), and on-load tap changers (OLTCs). As the operator of the active distribution network, the distribution company aims to minimize the total operating cost of the active distribution network.

[0023] (2) Due to its dispersion and the limitation of the power generation scale, it is difficult for distributed power sources on the user side to participate in the market alone. Therefore, load aggregators are proposed to combine multiple uncontrollable distributed power sources and some adjustable loads together to form a market-competitive entity to participate in electricity trading. The Load Aggregator Alliance (LAA) is responsible for the centralized management and coordination of all LA electricity demands, aiming to minimize the total operating cost of all LAs.

[0024] (3) Distributed energy storage participates in operation in the form of an operator. The operator participates in the competition by coordinating and controlling the charging and discharging power of each distributed energy storage unit in each time period and formulating the discharge price, aiming to minimize the total operating cost of all energy storage.

[0025] The operation of the multi-agent trading model can be described as follows: (1) When there is an electricity surplus or deficit within the LAA, to ensure economic benefits, LAs give priority to electricity trading within the alliance. LAs negotiate with each other to determine the trading volume and price of electricity, and realize the transfer of electricity through the distribution network. The two trading parties share the transmission fees charged by the distribution company equally.

[0026] (2) When the electricity trading within the alliance cannot meet the electricity demands of LAs, the LAA will conduct electricity interaction with the energy storage operator to achieve power balance through the charging and discharging of distributed energy storage. The trading electricity price and volume are determined through negotiation between the two trading parties, and the transmission fees are shared.

[0027] (3) When the above two trading forms still cannot meet the electricity demands of the load aggregator, the LA can conduct electricity trading with the DSO at a fixed time-of-use electricity price. In addition, to pursue higher profits, the energy storage operator can also conduct electricity trading with the distribution company at a fixed time-of-use electricity price.

[0028] (4) To meet the load demand and improve the operation efficiency, the distribution company can participate in the wholesale market and conduct electricity trading with the superior power grid.

[0029] Step S2: According to the multi-agent collaborative trading architecture, based on the Nash bargaining theory, construct a multi-agent cooperative game model, that is, construct an ADN collaborative optimal scheduling model considering the balance of distributed energy storage SOC and multi-agent cooperative game.

[0030] Step S2.1: Based on the multi-agent collaborative trading architecture, with the goal of minimizing costs and combined with constraint conditions, establish the operation and revenue models for each agent. Among them, as one of the agents, the distributed energy storage operator uses the distributed energy storage allocation strategy based on the SOC balance of distributed energy storage to constrain the operation of distributed energy storage.

[0031] In a multi-agent system, DSO, DESO, LAA, and each LA are all independent and rational individuals. Each agent will not interact with other agents regardless of its own interests. Therefore, the interest relationships among agents are both competitive and cooperative, with complex game relationships. Therefore, when agents participate in cooperative operation, they not only focus on the increase of their own interests, but more importantly, on the fair and reasonable determination of interest distribution. As an important branch of cooperative game theory, the Nash bargaining theory can be used to describe the cooperative interaction among multiple participants, with collective rationality and social optimization as the core. For a bargaining problem, it includes a series of agreements and disagreements. The bargaining disagreement point usually refers to the plan adopted in the most unfavorable scenario. In this embodiment, the operation result when each agent only has power interaction with the upper-level network is selected as the bargaining disagreement point. On this basis, for each agent, with the goal of minimizing costs, establish constraint conditions, and build the operation and revenue models for the distribution system operator DSO, the load aggregator LA, and the distributed energy storage operator DESO.

[0032] (1) DSO operation and revenue model With the goal of minimizing the total operating cost, and with power balance constraints, node injection power constraints, node voltage and branch current constraints, gateway power constraints, and power trading unidirectional constraints as constraint conditions, build the operation and revenue model for the distribution system operator DSO. Specifically: (1.1) Objective function For DSO, regardless of whether the bargaining is successful or failed, its overall goal is to minimize the total operating cost, which can be expressed as: (1) (2) In the formula, the superscripts Ag and Dag respectively represent the situations when an agreement is reached and a disagreement occurs during the bargaining process; WM and RM respectively represent the wholesale market and the retail market; t is the scheduling time interval; T is the optimization period; / and / are respectively the wholesale market trading cost and the retail market trading cost of DSO in the case of bargaining failure / success; are respectively the total costs of DSO in the case of bargaining failure / success; It is the network usage fee charged by the DSO to both the buyer and the seller.

[0033] Furthermore, the above-mentioned "cost" is not a single operating cost in the traditional sense, but takes into account a wider range of costs and revenues, including the DSO's direct power purchase cost and power sales revenue in the wholesale market, the transaction volume in the retail market, and the network usage fee charged during the peer-to-peer market transaction process. Among them, when the transaction volume is positive, it represents an expenditure, while a negative value represents revenue. All expressions of "cost" in this embodiment conform to this definition.

[0034] Among them, the wholesale market transaction cost is: (3) (4) In the formula, / is the power purchase / sale price of the DSO for power transactions with the upper-level power grid in the t period; / is the power quantity of the DSO interacting with the upper-level power grid in the case of bargaining failure / success; the operator means .

[0035] The retail market transaction cost is: (5) In the formula, / is the power price of the DSO purchasing / selling power from the LAA and DESO in the retail market in the t period; is the set of nodes where traditional power users are located; is the power quantity of the i-th aggregator (i.e., load aggregator #i) trading in the retail market in the case of successful bargaining; / are the charging / discharging powers of the DESO in the retail market in the case of successful bargaining, respectively; is the power consumption of the fixed load.

[0036] The network usage fee is: (6) In the formula, is the set of nodes of all other load aggregators excluding load aggregator # i ; is the set of nodes of all load aggregators; represents the set of nodes where energy storage is located; is the power transmission cost coefficient; the operator means taking the absolute value of the variable ; is for load aggregator #i To the load aggregator# j The amount of electricity exchanged, with a positive value indicating purchase and a negative value indicating sale; / The charging / discharging power provided by the energy storage for the load aggregator# i

[0037] (1.2) Constraints (1.2.1) Power balance constraint According to the network topology of the radial distribution network, the Distflow form of the power flow equation is adopted in this embodiment, which is: For node j it can be expressed as: (7) (8) For branch ij it can be expressed as: (9) (10) In the formula, represents the set of the head nodes of the branches with j as the end nodes in the power grid; represents the set of the end nodes of the branches with j as the head nodes in the power grid; and are t the three-phase active power and reactive power at the first end of the branch ij at time represents t the current of the branch ij at time and are respectively t the active power and reactive power of node j at time , are t the three-phase voltage amplitudes of node i, j at time and are ij the three-phase branch resistance and reactance of the branch.

[0038] (1.2.2) Node injection power constraint (11) (12) In the formula, and are respectively​t Time node j The active power and reactive power and are respectively t the active power and reactive power of the DG connected to node j at time and are respectively t the active power and reactive power of the load connected to node j at time is the reactive power of the reactive power compensation device connected to node j at time

[0039] (1.2.3) Node voltage and branch current constraints (13) In the formula, is t the voltage amplitude of each phase at node i at time and are respectively the maximum and minimum limits of the voltage amplitude, is t the current amplitude of branch ij at time is the critical current amplitude of branch overload

[0040] (1.2.4) Tie-line power constraints To suppress the impact of the power fluctuation of the ADN on the transmission network, it is necessary to consider the power constraints of the transmission network interface switch on the root bus of the distribution network, which can be expressed as: (14) In the formula, is the active power flowing from the transmission network interface into the distribution network at t time and are the upper and lower limits of the active power exchange of each relevant port set by the control center. The reactive power constraint is similar, that is is the reactive power flowing from the transmission network interface into the distribution network at t time and are the upper and lower limits of the reactive power exchange of each relevant port set by the control center

[0041] (1.2.5) Unidirectional power trading constraints To prevent LA from making a profit by buying electricity at a low price and selling it at a high price, its power trading is restricted, and it is not allowed for prosumers to conduct two-way power buying and selling transactions at the same time, which can be expressed as: (15) where M is a very large positive integer; is a 0-1 variable representing the load aggregator # i at time period t net load status, indicating that the net load of the load aggregator # i is positive, i.e., the local power generation cannot meet the load demand. At this time, it is only allowed to purchase electricity from other load aggregators, energy storage, or DSO, and selling electricity is prohibited, and vice versa; is the electricity quantity traded by the load aggregator #i in the retail market under the successful bargaining scenario; is the load aggregator # i to the load aggregator # j exchanged electricity quantity, with a positive value indicating purchase and a negative value indicating sale; / is the charging / discharging power provided by the energy storage for the load aggregator # i

[0042] (2) LA operation and revenue model Aiming at minimizing the total transaction cost and subject to the constraints of distributed renewable energy output, interruptible load, P2P transaction, and local energy balance, a load aggregator LA operation and revenue model is constructed. Specifically: (2.1) Objective function For the i th LA (i.e., the load aggregator LA# i ), in the case of negotiation failure and negotiation success, the overall objective is: (16) (17) In the negotiation failure scenario, the load aggregator # i only conducts electricity transactions with the DSO in the retail market, is the transaction cost of the load aggregator # i in this case; in the negotiation success scenario, the load aggregator # i can also conduct electricity transactions with other load aggregators and DESO in the P2P market, is the total transaction cost of the load aggregator # i , is the transaction cost for participating in the P2P market; in the negotiation success scenario, the load aggregator # i only conducts electricity transactions with the DSO in the retail market, is the transaction cost of the load aggregator # i in this case.

[0043] ​Among them, the retail market transaction cost is: (18) In the formula, / is the electricity price for the electricity purchased / sold by DSO from LAA and DESO in the retail market during period t; is the electricity quantity traded by load aggregator #i in the retail market under the successful bargaining scenario.

[0044] The P2P transaction cost is: (19) In the formula, the P2P market transaction cost consists of and The former refers to the fees paid for electricity trading within the load aggregator alliance and the network usage fees, and the latter refers to the costs and network usage fees paid for electricity trading with DESO. / is the electricity trading fee paid by load aggregator # i to load aggregator # j / DESO# s A positive value indicates buying, and a negative value indicates selling; is the adjustment parameter.

[0045] (2.2) Constraint conditions (2.2.1) Distributed renewable energy output constraint In the failed bargaining scenario, the load aggregator will increase the active power output of renewable energy as much as possible to meet its own electricity demand, and at the same time sell the surplus electricity to obtain profits, which can be expressed as: (20) In the formula, is the upper limit of the active power that can be output by renewable energy in load aggregator # i ; is the planned active power output of renewable energy in load aggregator # i at time t in the failed bargaining scenario.

[0046] In the successful bargaining scenario, the load aggregator can cooperate with DSO to provide reactive power to regulate the system power flow, which can be expressed as: (21) In the formula, and are the planned active and reactive power outputs of renewable energy in load aggregator # i in the successful bargaining scenario; is load aggregator #i The rated power of the renewable energy source.

[0047] Furthermore, the renewable energy sources considered in this embodiment include two types: distributed photovoltaic power generation devices and distributed wind turbines. Since their output characteristics are similar and their operating modes can be described by the above mathematical formulas, they are not distinguished here.

[0048] (2.2.2) Interruptible load constraint (22) In the formula, and are the upper and lower limits of the interruptible power of the interruptible load, represents the interruptible load; represents the number of interruptions of the interruptible load (IL) agreed in the contract, represents the maximum number of interruptions of the interruptible load (IL) agreed in the contract; represents the continuous response time of the interruptible load, and are the upper and lower limits of the continuous interruption time.

[0049] (2.2.3) P2P trading constraint During the P2P trading process, the electricity price and electricity quantity of both trading parties should satisfy: (23) (24) In the formula, / is the electricity trading fee paid by the load aggregator # i to the load aggregator # j / DESO# s Positive values indicate buying, and negative values indicate selling; is the electricity quantity exchanged by the load aggregator # i to the load aggregator # j Positive values indicate buying, and negative values indicate selling.

[0050] (2.2.4) Local energy balance constraint For each load aggregator, the power flow should be balanced at any time. Therefore, the total trading electricity quantity in each trading market should be equal to the local net load, which can be expressed as: (25) (26) In the formula, is, in the case of a successful negotiation scenario, the load aggregator # i at time periodt Net load prediction.

[0051] (3) DESO Operation and Revenue Model Aiming at maximizing revenue and taking the constructed operation constraints of distributed energy storage as constraints, an operation and revenue model for distributed energy storage operator DESO is constructed. Specifically: (3.1) Objective Function In the cases of successful and failed bargaining, the goal of DESO is to maximize revenue, which is equivalent to minimizing cost, specifically expressed as: (27) (28) In the formula, / is the cost of DESO's electricity transaction with DSO in the retail market in the case of failed / successful bargaining t during the period; is the cost of DESO's transaction with prosumers in the P2P market in the case of successful bargaining; / is the degradation cost of shared energy storage charging and discharging in the case of failed / successful bargaining.

[0052] Among them, the retail market cost is: (29) The P2P transaction cost is: (30) The shared energy storage degradation cost is: (31) (32) In the formula, is the cost coefficient of DES life loss, which is a comprehensive index used to describe the performance degradation and loss of DES during charging and discharging.

[0053] (3.2) Constraints In this embodiment, considering the problem of unbalanced state of charge (SOC) of distributed energy storage (DES), a distributed energy storage allocation strategy based on DES SOC balance is proposed, and the operation constraints of distributed energy storage are constructed according to this strategy.

[0054] Specifically, the proposed distributed energy storage allocation strategy based on DES SOC balance is: First, divide several distributed energy storage units in the distributed energy storage system into two groups that respectively perform charge / discharge tasks, and determine the priority charge / discharge states of the two groups according to the actual SOC conditions of the distributed energy storage units in the two groups.

[0055] Specifically, in the actual scheduling process, in order to suppress the output fluctuations of renewable energy, the energy storage system needs to frequently switch the charge / discharge state, which to a certain extent increases the charge / discharge frequency of the energy storage, accelerates the aging of the battery, and at the same time the capacities of the individual energy storage units are not fully utilized. Based on the above considerations, the distributed energy storage units in the energy storage system are divided into two groups to respectively perform charge / discharge tasks, so as to reduce unnecessary charge / discharge state switching, thereby reducing the charge / discharge frequency of the battery and reducing the life loss.

[0056] Furthermore, for a distributed energy storage system composed of 2N distributed energy storages, at the initial moment, the N distributed energy storages with lower SOC are set as the priority charging group, and the N distributed energy storages with higher SOC are set as the priority discharging group. It should be noted that there are certain moments when the total power command exceeds the rated power of a single group, and cross-group coordination is required to provide short-term power support to jointly complete the total power command for this period. In addition, when the SOC balance within the group is poor or the SOC of any single unit reaches a predetermined threshold during operation, the role exchange between the groups will be triggered; the threshold can be adjusted adaptively according to the system operation. In the normal operation scenario, the goal of setting the threshold is to minimize the loss of the DES life; in extreme cases, the SOC limit will be appropriately relaxed to ensure the stable operation of the system.

[0057] Secondly, execute a two-layer power distribution strategy, including a power upper-layer distribution strategy and a power lower-layer distribution strategy, as follows: The power upper-layer distribution strategy, that is, based on the priority charge / discharge states of the two groups, determine the charge / discharge demand responses and response sequences of the two groups according to the actual charge / discharge demands of the distributed energy storage system. Among them, after determining the priority charge / discharge states of the two battery groups, the total charge / discharge command of the system needs to be allocated to the two battery groups. First, it is necessary to determine the response sequence of the two battery groups: If it is a charging command, that is, <0, the priority charging group responds first. On the contrary, >0, the priority discharging group responds first until the charge / discharge power meets the total command requirements or reaches the maximum power of the battery group. After the total power command is allocated by the upper layer, assuming that the maximum charge / discharge powers of the two battery groups are both , the charge / discharge powers borne by the two battery groups are shown in Table 1.

[0058] Table 1: Charge / discharge powers borne by the two battery groups

[0059] The power lower-layer distribution strategy is to distribute the power to each distributed energy storage unit in each group according to the charge / discharge demand response and response order of the two groups. Among them, the power distribution process is as follows: when the charge / discharge demand of the current group is the maximum value, the power is distributed according to the maximum charge / discharge power principle; otherwise, the power is distributed according to the SOC balance principle.

[0060] Specifically, after the upper layer obtains the charge / discharge instructions that the two battery packs should undertake, it is necessary to further determine the internal power distribution. When the charge / discharge instruction of the battery pack is , the lower layer distributes according to the maximum charge / discharge power principle, that is, the power instruction borne by each distributed energy storage is the maximum charge / discharge power ; when the charge / discharge instruction of the battery pack is not , the lower layer distributes according to the SOC balance principle, allowing the distributed energy storage with a higher SOC to charge less and discharge more, and the distributed energy storage with a lower SOC to charge more and discharge less, so as to achieve the relative balance of SOC among the distributed energy storages.

[0061] The above strategy can be described in detail by the following formula, which is: First, use the Sigmoid function to describe the charge / discharge function of the distributed energy storage, that is: (33) In the formula, is t the SOC value of the energy storage unit at i -1 moment, and respectively represent the charge function and discharge function of the distributed energy storage, m and n are the parameters affecting the function slope and center point.

[0062] Secondly, to ensure the reasonable distribution of power within the group, and considering the rated power and SOC value of each distributed energy storage to determine its output power, it can be expressed as: (34) In the formula, and are respectively the charge / discharge power borne by the energy storage i at t moment; is the rated charge / discharge power of the energy storage i ; N is the number of distributed energy storages.

[0063] Furthermore, the SOC balance distribution strategy based on the charge / discharge function is shown in the following formula: (35) In the formula, is the battery pack k At t the charging and discharging power that needs to be borne at the moment.

[0064] Furthermore, considering that the distributed energy storage should respond to the charging and discharging instructions on the premise of meeting the operation constraints, according to the above strategy, the operation constraints of the distributed energy storage DES are constructed as follows: (3.2.1) Charge / discharge constraint The charging and discharging power of the energy storage at any moment should not exceed its maximum charging and discharging power, which is: (36) (37) In the formula, and are the charging / discharging power of the DES at t the moment; and are the maximum charging / discharging power of the DES. Considering that the energy storage system can only be in the charging or discharging state at a certain moment, and there is no phenomenon of simultaneous charging and discharging, the 0 / 1 variables and are introduced. , indicating that the energy storage system is in the charging state; , indicating that the energy storage device is in the discharging state, and its constraints are expressed as follows: (38) (3.2.2) State of charge constraint In order to ensure that the energy storage system has the same regulation performance in the new scheduling period, the state of charge should satisfy: (39) Among them, is the remaining state of charge of the DES at time t; and are the charging / discharging efficiency of the DES; ∆t is the scheduling time interval.

[0065] (3.2.3) SOC constraint When the energy storage is charging, it absorbs active power and the SOC increases; when discharging, it emits active power and the SOC decreases. The calculation formula for the SOC value of the energy storage unit at time t is: (40) In the formula, is the SOC value at the initial moment, and are the charging and discharging powers respectively during the period from t to t - 1, and at most only one of them can be non-zero; and are the charge-discharge efficiency, both less than 1; is the duration of charge and discharge (scheduling period); is the rated capacity of the energy storage.

[0066] In addition, to improve the SOC balance degree, when the charge-discharge command k of the battery pack is not 0, the standard deviation of the SOC of the battery units in the battery pack k should gradually decrease; when is 0, the standard deviation of the SOC of the battery units in the battery pack k should remain unchanged. This constraint can be expressed as: (41) Step S2.2. According to the Nash bargaining theory, a multi-agent cooperation game model is constructed by integrating the operation and revenue models of multiple agents.

[0067] Based on the Nash bargaining theory, the above problem can be modeled as: (42) Step S3. Based on the actual energy demand on the user side, the multi-agent cooperation game model is solved to obtain the optimal scheduling strategy of energy in the active distribution network, and energy scheduling is carried out according to the scheduling strategy.

[0068] Specifically, in the original problem, both the energy trading volume and the energy trading price are used as optimization variables, and there is a product term between them. Essentially, it is a mixed integer non-convex optimization problem and is difficult to solve directly. To reduce the complexity of problem solving, the model is transformed into two linear sub-problems that are easy to solve: maximizing social benefits (denoted as sub-problem 1) and cooperative revenue distribution (denoted as sub-problem 2), and the optimal solution of the original problem is obtained through sequential optimization, where: For sub-problem 1: maximizing social benefits, it can be expressed as: (43) In the formula, and are the wholesale market transaction cost and the retail market transaction cost of the DSO in the successful bargaining scenario respectively, is the charge-discharge degradation cost of the shared energy storage of the DESO in the successful bargaining scenario.

[0069] By solving the above sub-problem 1, the variable values can be optimized: t The three-phase active power ij at the head node of branch and the reactive power at time ij the square of the current amplitude of branch at nodei The square of the voltage amplitude The electricity quantity purchased (positive value) or sold (negative value) by the DSO in the wholesale market under the successful negotiation scenario Under the successful negotiation scenario, the load aggregator #i The electricity quantity traded in the retail market The charging / discharging power of the DESO in the retail market under the successful negotiation scenario / The load aggregator # i The electricity quantity exchanged with the load aggregator # j The load aggregator # The charging / discharging power provided by the energy storage for the load aggregator # i / / Under the successful negotiation scenario, the load aggregator # i The planned active power output of renewable energy And the reactive power output .

[0070] For sub - problem 2: maximizing social benefits. Since the time - of - use electricity price at which the DSO trades with the upper - layer power grid considered in this embodiment is a fixed one, only the load aggregator alliance and the energy storage operator participate in the negotiation in sub - problem 2. The participants negotiate according to their respective bargaining powers, and finally determine an energy trading price that satisfies all bargaining parties. Specifically: First, construct the first - stage cooperative revenue distribution model according to the bargaining powers of the load aggregator alliance and the energy storage operator, which is: (44) Furthermore, based on the obtained revenue of the load aggregator alliance , redistribute the revenue among the load aggregators. Construct the second - stage cooperative revenue distribution model according to the bargaining powers of the load aggregators, which is: (45) Finally, take the logarithm of the formula of the above - mentioned second - stage cooperative revenue distribution model, and then take its opposite number to transform the product - maximum problem into a sum - minimum problem, which is: (46) (47) Substitute the energy trading volume obtained from sub - problem 1 into the above equation. By solving the equation, the cooperative benefit distribution is obtained, and thus the energy trading price is obtained. That is, at this time, the optimal scheduling strategy of energy in the active distribution network is obtained, and then the corresponding energy scheduling can be carried out according to this scheduling strategy.

[0071] Furthermore, through the case study of the IEEE 33-node system, the effectiveness of the method proposed in this embodiment is further verified.

[0072] Specifically, first, a case is set up. Taking the improved IEEE 33-node distribution system as an example, the effectiveness of the proposed method is verified. This model divides the 24-hour scheduling cycle into 24 time periods, sets the node voltage operation limit values to 0.95 p.u. and 1.05 p.u., and sets the operating costs of wind energy and photovoltaic to 0 to promote the consumption of renewable energy.

[0073] The case study system includes 4 load aggregators (nodes 13, 16, 24, and 29) and 6 distributed energy storages (nodes 3, 10, 15, 21, 30, 31). For ease of description, the 4 load aggregators are named load aggregator #1 - load aggregator #4 in ascending order of node numbers. Load aggregator #1 and load aggregator #3 are mainly powered by photovoltaic power generation, and load aggregator #2 and load aggregator #4 are mainly powered by wind power generation. The scheduling time interval is 1 h, the optimization period is 24 h, and the power transactions between each market participant and the DSO adopt a fixed time-of-use electricity price. The electricity purchase and sale price of the DSO from / to the upper-level power grid is set to 0.8 times the electricity price for transactions with other market participants. The modeling tool for case study simulation is Yalmip, the operating environment is MATLAB 2020a, the solver called is Cplex, and the computer hardware configuration is Intel(R) Core(TM) i5-11400 2.60GHz, 16G RAM.

[0074] To verify the effectiveness of the method proposed in this embodiment, 3 scenarios are set up for comparative analysis: Case1: Without considering the cooperation among multiple entities, LLA and DESO only trade with the DSO.

[0075] Case2: Without considering the cooperation among multiple entities, but adopting a distributed energy storage scheduling strategy based on SOC balance.

[0076] Case3: The proposed method.

[0077] Secondly, scenario generation and reduction. Generally, the probability density distribution function of the prediction error of wind and light output satisfies the normal distribution, but the actual probability distribution is unknown. If the normal distribution is directly adopted, a large error will occur. Therefore, in this embodiment, based on historical data, the kernel density estimation method is used to obtain the actual probability density distribution function of the prediction error of renewable energy output at each scheduling moment, and random scenario generation and reduction are carried out. Finally, 10 typical wind output scenarios are obtained as follows Figure 3 as shown.

[0078] After that, analyze the operation results of distributed energy storage. The daily change curves of the distributed energy storage SOC under the distributed energy storage SOC balancing scheduling strategy and without adopting the proposed strategy are shown in Figure 4 and Figure 5 respectively. It can be seen from the comparison of the figures that without the balanced strategy scheduling, the SOC balance degree among energy storages is poor, and the utilization rate of energy storage capacity is also low. When the SOC balancing strategy is adopted to schedule the distributed energy storage, although the initial SOC of each energy storage unit is different, with the passage of time, the gap gradually narrows, and the SOC values of each energy storage in the group slowly converge and tend to be consistent. When the state of charge of the energy storage in the group reaches the upper and lower limits, the two groups will exchange "roles", which can effectively improve the utilization rate of energy storage capacity.

[0079] Further analyze the economy. To verify the effectiveness of the method proposed in this embodiment in improving the economic benefits of each market entity, optimize and solve Scheme 1, 2, and 3 respectively. The comparison of the total costs of each entity under the three schemes is shown in Table 2 below. A positive cost indicates expenditure, and a negative cost indicates income.

[0080] Table 2: Comparison of the costs of each entity in three cases

[0081] When considering the cooperative game and successful bargaining scenarios among multiple entities, the overall economic efficiency of the system is significantly improved. For the distribution system operator, the total dispatching cost is greatly reduced. This is because when multiple entities cooperate, they can jointly optimize the power flow of the distribution system, promote the maximized utilization of local energy, increase the consumption of renewable energy, and thus reduce the cost of the distribution network purchasing electricity from the superior grid. Among them, the total costs of the DSO in Scheme 2 and 3 are equal because the DSO trades at a fixed electricity price in each market. When the energy trading volume is determined, its income is also determined, that is, the income of the DSO in Scheme 3 is not redistributed.

[0082] Compared with Scheme 1, the costs of the DESO and the load aggregator in Scheme 3 are reduced to varying degrees. For the DESO and the load aggregator, they can participate in the retail market and the P2P market at the same time. The electricity price in the retail market is fixed, and the selling electricity price of the DSO is higher than the purchasing electricity price. Therefore, in the trading process, they can more flexibly select favorable trading partners for themselves to increase income.

[0083] Table 3: Benefit distribution of Scheme 3

[0084] Further analyze the dispatching results. To test the effectiveness of the proposed method in optimizing the power flow of the distribution system and improving the system operation status, Figure 6It shows the comparison of the peak shaving and valley filling effects of the system under three scenarios (i.e., Method1 - 3).

[0085] As can be seen from Figure 6

[0086] Compared with Method 1 and 2, the load curve of the method proposed in this embodiment is smoother. This is because on the one hand, distributed energy storage can trade with the DSO and cooperate with the DSO to optimize the power flow; on the other hand, it can store the excess power of the load aggregator and release electrical energy to the load aggregator in a power - shortage state. At the same time, using its energy time - shifting characteristic, distributed energy storage transports electricity among load aggregators, making its total net load curve more stable, thus alleviating the power supply pressure caused by the large peak - valley difference of the distribution network load.Embodiment 2 This embodiment provides an active distribution network optimal scheduling system considering multi - agent cooperative game, including: An optimal scheduling model construction module, which is used to consider the interaction among multiple market agents in the active distribution network, construct a multi - agent collaborative trading architecture; based on the multi - agent collaborative trading architecture, with the goal of minimizing cost and combining constraint conditions, build the operation and revenue models of each agent; among them, as one of the agents, the distributed energy storage operator uses a distributed energy storage allocation strategy based on the SOC balance of distributed energy storage to constrain the operation of distributed energy storage; according to the Nash bargaining theory, integrate the operation and revenue models of multiple agents to construct a multi - agent cooperative game model; A model solving and scheduling module, which is used to solve the multi - agent cooperative game model based on the actual energy demand on the user side, obtain the optimal scheduling strategy of energy in the active distribution network, and perform energy scheduling according to the scheduling strategy.

[0087] Embodiment 3 This embodiment provides an electronic device, including: a memory for storing executable instructions; a processor for implementing the above - mentioned method provided by this embodiment when executing the executable instructions stored in the memory.

[0088] Embodiment 4 This embodiment also provides a computer - readable storage medium storing executable instructions, which will cause the processor to execute the above - mentioned method provided by this embodiment when the executable instructions are executed by the processor.

[0089] Embodiment 5 This embodiment provides a computer program product, which includes executable instructions, and the executable instructions are a kind of computer instructions; the executable instructions are stored in a computer - readable storage medium. When the processor of an electronic device reads the executable instructions from the computer - readable storage medium and executes the executable instructions, the electronic device is caused to execute the above - mentioned method provided by this embodiment.

[0090] The steps involved in the second to fifth embodiments above correspond to those in the first method embodiment. For specific implementation manners, reference may be made to the relevant description part of the first embodiment. The term "computer-readable storage medium" should be understood to include a single medium or multiple media containing one or more instruction sets; it should also be understood to include any medium that can store, encode, or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0091] Those skilled in the art should understand that the above-mentioned modules or steps of the present invention can be implemented by a general-purpose computer device. Optionally, they can be implemented by program codes executable by a computing device, so that they can be stored in a storage device for execution by the computing device, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0092] The above are only the preferred embodiments of the present invention. Although the specific implementation manners of the present invention have been described in conjunction with the accompanying drawings, it is not a limitation on the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solutions of the present invention, various modifications or deformations that can be made without creative efforts by those skilled in the art are still within the protection scope of the present invention.

Claims

1. An active distribution network optimization dispatching method considering multi-agent cooperative game, characterized in that: include: Considering the interaction between multiple market players in the active distribution network, a multi-player collaborative transaction framework is constructed; Based on a multi-agent collaborative transaction architecture, with the goal of minimizing costs and combined with constraints, an operation and revenue model for each entity is built; among them, the distributed energy storage operator, as one of the entities, uses a distributed energy storage allocation strategy based on distributed energy storage SOC balance to constrain the operation of distributed energy storage; According to Nash bargaining theory, the operation and benefit models of multiple subjects are integrated to construct a multi-subject cooperative game model; Based on the actual energy demand on the user side, the multi-agent cooperative game model is solved to obtain the optimal scheduling strategy for energy in the active distribution network, and energy scheduling is carried out according to the scheduling strategy.

2. The active distribution network optimization dispatching method considering multi-agent cooperative game as claimed in claim 1 is characterized in that: The distributed energy storage allocation strategy based on distributed energy storage SOC balancing is: Divide a number of distributed energy storage units in a distributed energy storage system into two groups that perform charging / discharging tasks respectively, and determine the priority charging / discharging states of the two groups according to the actual SOC conditions of the distributed energy storage units in the two groups; Based on the priority charge / discharge states of the two groups and the actual charge / discharge demands of the distributed energy storage system, the charge / discharge demand responses and response order of the two groups are determined; According to the charging / discharging demand responses and response order of the two groups, power is allocated to each distributed energy storage unit in each group; the power allocation process is as follows: If the charge / discharge demand of the current group is the maximum value, power allocation is performed according to the maximum charge / discharge power principle; otherwise, power allocation is performed according to the SOC balance principle.

3. The active distribution network optimization dispatching method considering multi-agent cooperative game as claimed in claim 1 is characterized in that: The multiple entities include distribution system operator DSO, load aggregator LA and distributed energy storage operator DESO; for each entity, constraints are established with the goal of minimizing costs, and operation and revenue models of distribution system operator DSO, load aggregator LA and distributed energy storage operator DESO are built; Among them, with the goal of minimizing the total operating cost, the distribution system operator DSO operation and revenue model is constructed with power balance constraints, node injection power constraints, node voltage and branch current constraints, gateway power constraints, and power trading unidirectional constraints as constraints; among them, the total operating cost of the distribution system operator includes wholesale market transaction costs, retail market transaction costs and network usage fees.

4. The active distribution network optimization dispatching method considering multi-agent cooperative game as claimed in claim 3 is characterized in that: With the goal of minimizing the total transaction cost, the operation and profit model of the load aggregator LA is constructed with the distributed renewable energy output constraint, interruptible load constraint, P2P transaction constraint and local energy balance constraint as constraints; among which, the total transaction cost includes the retail market transaction cost and P2P transaction cost.

5. The active distribution network optimization dispatching method considering multi-agent cooperative game as claimed in claim 3 is characterized in that: According to the distributed energy storage allocation strategy based on distributed energy storage SOC balance, the operation constraints of distributed energy storage are constructed, which include: charging / discharging constraints, power constraints and SOC constraints; With the goal of maximizing profits and the constructed distributed energy storage operation constraints as constraints, the operation and profit model of the distributed energy storage operator DESO is constructed; wherein maximizing profits is equivalent to minimizing costs, which include retail market costs, P2P transaction costs, and shared energy storage degradation costs.

6. The active distribution network optimization dispatching method considering multi-agent cooperative game as claimed in claim 1 is characterized in that: In the process of solving the multi-agent cooperative game model, the model is converted into a two-linear sub-problem, and the optimal solution is obtained through sequential optimization, including: Based on the multi-agent cooperative game model, two sub-problems are separated: maximizing social benefits and allocating cooperative benefits. Solve the sub-problem of maximizing social benefits and obtain the energy trading volume; For the cooperative profit distribution sub-problem, the first-stage cooperative profit distribution model is first constructed according to the bargaining power of the load aggregator alliance and the energy storage operator. Then, based on the solved load aggregator alliance profit, the profit is redistributed among the load aggregators. The second-stage cooperative profit distribution model is constructed according to the bargaining power of each load aggregator. The second-stage cooperative profit distribution model is converted into a minimum value problem, and the energy trading volume is substituted to obtain the cooperative benefit distribution and energy trading price.

7. An active distribution network optimization dispatching system considering multi-agent cooperative game, characterized in that: include: The optimization dispatch model construction module is used to consider the interaction between multiple market players in the active distribution network and build a multi-player collaborative transaction architecture; Based on the multi-agent collaborative transaction framework, with the goal of minimizing costs and combined with constraints, the operation and profit model of each subject is built; among them, the distributed energy storage operator, as one of the subjects, uses the distributed energy storage allocation strategy based on the distributed energy storage SOC equilibrium to constrain the operation of distributed energy storage; according to the Nash bargaining theory, the operation and profit models of multiple subjects are integrated to build a multi-agent cooperative game model; The model solving and scheduling module is used to solve the multi-agent cooperative game model based on the actual energy demand on the user side, obtain the optimal scheduling strategy for energy in the active distribution network, and perform energy scheduling according to the scheduling strategy.

8. An electronic device, characterized in that: include: A memory for storing executable instructions; The processor is used to implement the active distribution network optimization scheduling method considering multi-agent cooperative game as described in any one of claims 1 to 6 when executing the executable instructions stored in the memory.

9. A computer-readable storage medium, characterized in that: Executable instructions are stored, which are used to cause the processor to execute the executable instructions to implement the active distribution network optimization scheduling method considering multi-agent cooperative game as described in any one of claims 1-6.

10. A computer program product, characterized in that The computer program product includes executable instructions stored in a computer-readable storage medium; When the processor of the electronic device reads the executable instructions from the computer-readable storage medium and executes the executable instructions, the active distribution network optimization scheduling method considering multi-agent cooperative game as described in any one of claims 1 to 6 is implemented.

Images